Multilingual capabilities of GPT: A study of structural ambiguity.

PLoS One · 2025 · PMC12233270 · PMID 40622933

Europe PMC DOI

hedged sentences

0.1000

closest p · 2.0× alpha

0.4000

boldest claim

The sentences

marginal significancep = 0.10so close (0.05 < p ≤ 0.1)

Statistical analysis confirmed a significant preference for LA over HA in Korean, with marginal significance in GPT-3.5-turbo (z = −1.63, p = 0.10).

also in 5,069 other papers

did not reach statistical significancep = 0.40not close (p > 0.1)

Question type B in GPT −4-turbo exhibited the same trend, favoring LA interpretations, but did not reach statistical significance (z = −0.83, p = 0.40).

also in 111,027 other papers

a numerical trendno p-value reported

Although the overall LA interpretation rate of o1-mini did not reach 60%, it was close—reaching 58.49%—and still showed a numerical trend toward LA preference.

also in 1,012 other papers

Quoted from the open-access full text in Europe PMC under the licence the publisher applied. The sentence is reproduced exactly as published; the emphasis is ours.